In-Memory Database (Multi-Part)
Hard · Hashing · about 60 minutes
Asked at Anthropic, Block (Square), Databricks, MongoDB, OpenAI, Oracle, Salesforce, Snowflake, Stripe, ServiceNow, Workday, Intuit, Visa, Capital One, Plaid, Ramp, Datadog, Confluent, Perplexity, HubSpot, Elastic, Chime, Affirm, Klarna, Revolut, American Express, xAI, Cohere, Hugging Face, Toast, Mistral AI, Dell Technologies, Nutanix, Rippling, Gusto, Brex, Mercury, Kraken, Akamai.
Build an in-memory database in stages. Level 1: SET/GET/DELETE on records with fields. Level 2: scan fields of a record, with prefix filtering. Level 3: TTLs on fields with a timestamp passed to every operation. Level 4: BACKUP and RESTORE to a point in time. Each level must keep earlier tests passing.